Multi-modal Discrete Collaborative Filtering

With the development of E-commerce, recommender systems have been widely adopted by many online services for helping their customers find desirable products to purchase. However, the ever-growing scales of products and users render recommendation more challenging than ever before [1]. For example, there are more than 0.82 billion active Taobao users and over one billion products for sale now. Consequently, it is challenging to make the immediate response to match products for potential customers accurately and efficiently, by analyzing large-scale yet sparse user interaction history.